AI Chatbot Plugin for Jenkins – GSoC 2025 Project Kickoff
Giovanni Vaccarino
Hi everyone,
I'm Giovanni Vaccarino, a GSoC 2025 contributor working with a Jenkins project this summer. I'm currently pursuing a double MSc in Computer Science and Engineering at Politecnico di Milano and the University of Illinois Chicago, with a focus on machine learning and natural language processing.
My project is titled "AI-Powered Chatbot for Quick Access to Jenkins Resources." The goal is to develop a Jenkins plugin that embeds a conversational assistant directly in the Jenkins UI. It will help users quickly locate relevant documentation, plugin information, and community-related resources.
The chatbot uses a Retrieval-Augmented Generation (RAG) approach to search across the collected data. It’s backed by a quantized local LLM served via FastAPI, with a React/TypeScript frontend for a modern UI.
During the bonding period, I:
- Participated in onboarding and weekly project meetings
- Set up the development environment and forked the plugin repository
- Implemented the core of the RAG pipeline and initiated the frontend/backend integration
- Merged early PRs on data collection, preprocessing and chunking
- Submitted WIP PRs for UI, and API integration
- Maintained a detailed work diary and conducted research to optimize RAG performance
Repo: https://github.com/jenkinsci/resources-ai-chatbot-plugin
Over the summer, I aim to:
- Improve the chatbot’s search quality and response accuracy
- Enhance the UI/UX for better usability within Jenkins
- Introduce evaluation strategies like LLM-as-a-Judge to evaluate the quality of the retrieval
- Add unit and integration tests to ensure correctness and robustness of the code
I'm excited to keep building and learning with the Jenkins community. Feedback, ideas, or suggestions are always welcome!
You can reach me out on:
- Linkedin: https://www.linkedin.com/in/giovanni-vaccarino-931686258/
- GitHub: https://github.com/giovanni-vaccarino
Best regards,
Giovanni Vaccarino